Technical development

BCI algorithms

Non-invasive Brain Computer Interface (BCIs) are device that allow a person to interact with the environment using no muscular contraction. At the state of the art several BCI system have been developed in a wide number of laboratories and have been demonstrated to provide for a valuable alternative communication channel in severely disabled people. However, their usage in everyday life is still limited mainly due to a system design that is addressed by developers within a research laboratory and therefore, when systems are moved into other locations (e.g. a patient home) the support from highly specialized technical staff, is missing or cannot be guaranteed. For instance, calibration procedures, even if not so complicated, are often required to prevent systems from becoming unusable. To tackle this issue, a special BCI device will be designed and realized by exploiting well defined and widely adopted techniques; these latter will be integrated into a special package in order to dramatically reduce the need for any highly specialized technical support. An embedded device will be developed to provide for the one side, a more reliable solution compared to PC-based one and on the other, a practical, portable BCI system. Built-in self calibration and tuning procedures will be also implemented in order to guarantee an optimal functioning of the BCI device without the need of any technician.

Gateway - Control to appliances routing

The Gateway module represents an hardware component acting as an interface between multiple modalities of user input and multiple available output devices. In fact, depending on personal preferences and on the level of motor abilities, the user may choose or need to utilize a specific input device (buttons, touchpad, switches,BCI , etc.); the Gateway would function as a common interface through which the user can control several output devices (e.g. a communication aid, a PC, a TV set, a domotic controller, etc). This approach has several advantages: (i) it provides a better match to the user's preferences, and (ii) it adapts to the gradual loss of motor functions due to the progress of ALS; (iii) it allows a user to familiarize in advance with a new input device that he may need to master in the future, when loss of motor skills prevents him to keep on using his current input device. The Gateway should abstract common aspects of each input device it is designed for, and should have an internal representation of all possible commands that the output devices might accept. Moreover, it should provide the user with a (visual) interface which intuitively represents the set of available commands or communication options. In order to be effective, the Gateway must successfully solve the following tradeoffs:

  • input performance vs.universal accessibility: a smart abstraction of input modalities should avoid flattening functionalities and ease of use toward the least complex.

  • output richness vs. universal control: a smart abstraction of output devices should avoid implementing the subset of common commands alone

  • control intuitiveness vs. portability: a smart user interface should allow selecting the desired option in a minimum number of steps, even on a small display

Dedicated hardware

As mentioned in the BCI algorithms section the BCI device will be integrated in an embedded and portable electronic device (EPED) in order to:

  • acquire and process the EEG signal in real time, aimed at performing the translation of voluntary modification of EEG signals into commands (Brain-Computer Interface). A commercial, CE marked, portable amplification front-end system will be used (g.Tec gMobiLab), aiming to obtain a digital, amplified EEG signals. The output of this amplifier will be sent through a physical interface (Bluetooth or Serial) to the embedded device (EPED). This module will perform the signals processing operation (signals pre-processing, signals features extraction and signals features classification). As output, several actuator signals will be generated and will be sent to domotics system through an opportune transmission protocol.

    • convert user commands generated through the BCI and/or with other devices based on muscular operation (e.g. switches, touch sensors, head trackers) into control signals for the appliance they are aimed for (communication aid, TV sets, motorized mattress, domotic appliances). Input devices will be physically connected via jack plugs (buttons and switches) or usb ports (trackpad and head trackers) or even wirelessly (mice and BCI device). The system will feature a portable display for configuration and operation. Outputs will interface with remotely controlled appliances (e.g. TV sets, via infrared port), to domotic controllers (konnex socket) and to simple communicators (via serial port).

The design of the embedded system will be based on a custom approach: for a rapid prototyping, we will use an integrated board which includes a reconfigurable field programmable gate array (FPGA), a real-time processor and digital I/O. Probably, the choice will be a NI Single-Board RIO that is a low-cost embedded deployment solutions based on NI CompactRIO technology.